Apollo Research

Apollo Research is an AI safety organization specializing in model evaluations and red-teaming for frontier AI systems. Its systematic evaluation methodology tests AI models for dangerous capabilities including deceptive reasoning, power-seeking behaviors, and scheming to avoid oversight, providing independent assessments that inform deployment decisions by AI labs and regulators. Apollo Research's published evaluations of major frontier models have contributed to establishing standards for pre-deployment safety testing, and its work on detecting situational awareness in AI systems addresses one of the most challenging problems in AI safety evaluation.